Buffet holding time

Why can a hotel buffet dish that still feels comfortably warm be far more dangerous than one that has gone cold?

Two trays at the end of a hotel buffet. One holds rice gone stone cold and looking unappetising. The other holds a curry still pleasantly warm — warm enough that nothing about it says stop. Every instinct ranks the cold tray as the riskier one.

The instinct is backwards, and the reason is worth taking slowly. Our sense of danger in food runs on appetite: cold and congealed reads as spoiled. But the organisms that make you ill are not the organisms that make food look bad, and neither cares what your hand thinks. So what is the tray's temperature actually telling us?

Begin with the only thing that really matters, a growth rate. Bacteria divide, and division is multiplicative. Suppose a dish leaves the kitchen carrying a small population of survivors and sits out for six hours. If those organisms double every twenty minutes, six hours is eighteen doublings, and two to the eighteenth is about 260,000 — one cell becomes a quarter of a million. Now the same six hours at a ninety-minute doubling time: four doublings, sixteenfold. Same food, same time, same starting population, four orders of magnitude between the outcomes. That is the mechanism in one calculation; the rest is detail about what sets the doubling time.

Temperature sets it, and not gently. Growth stops at the cold end because the chemistry slows and at the hot end because proteins come apart. Between, most food-poisoning bacteria have an optimum around body heat and a little above — broadly the mid-thirties to mid-forties Celsius, though the peak varies by organism, and I am recalling that range rather than deriving it. Regulators bracket this as a "danger zone": the hot-holding floor is 63 °C in United Kingdom practice and 57 °C under the United States Food Code, with a chilled ceiling of 5 °C in most places and 8 °C in some. All recalled, and they genuinely differ by jurisdiction — itself a hint that they are conservative brackets rather than sharp physical edges.

Now place the two trays on that scale. Properly chilled rice sits below the lower bracket, where doubling times run to many hours. The warm curry, at the hand-pleasant temperature of forty-odd degrees, sits almost exactly on the optimum. It is not a little worse; it is the worst temperature available, and it feels fine precisely because it is close to the temperature of your own hand.

But that raises an awkward question. The curry was cooked, and cooking kills the vegetative bacteria that arrived on the raw ingredients. Where is the doubling population coming from?

Here the tidy story about germs landing on food runs out, and the real answer is more specific. Some bacteria — notably Clostridium perfringens and Bacillus cereus — form spores, dormant structures that survive normal cooking comfortably. Cooking therefore does two things at once: it fails to remove the spores, and it removes everything that would have competed with them. When the dish cools through the zone, those spores germinate into an environment that is warm, moist, nutritious and swept clean of rivals. Cooking has not sterilised the dish; it has selected the one group that can exploit it, and prepared the medium.

That is why outbreaks from food held warm in bulk involve these organisms so consistently, rather than the Salmonella and Campylobacter that dominate raw-food illness. And C. perfringens is fast: generation times of the order of ten minutes at its optimum are reported, which shortens the arithmetic above alarmingly.

One more turn, and it defeats the obvious remedy. If the problem were living cells, reheating would solve it. But B. cereus produces cereulide, an emetic toxin that is heat-stable and survives reheating that kills every cell in the pan. Part of what accumulated during those hours cannot be undone by heat at all.

"Reheat it thoroughly and it is fine" is the most common and most dangerous folk wisdom here, and the toxin case refutes it. So does "you would smell it": spoilage organisms and pathogens are different populations with different requirements, and a dish carrying a heavy pathogenic load can taste entirely normal. Neither nose nor palate is an instrument.

Nor is this best told as a story of negligent staff, because much of the cause sits in the equipment. A wide, shallow chafing dish over a gel burner in an air-conditioned room is often simply unable to hold 63 °C at the surface, whatever the crew intends. Topping up a half-empty tray with fresh food is a structural trap of the same kind: it resets the temperature but not the clock, because the old food underneath keeps its accumulated hours.

The honest limit: those brackets are policy simplifications of a continuous curve, drawn with a margin. Nothing dramatic happens at 5.0 °C, and a dish at 8 °C is not equivalent to one at 40 °C.

And what would refute the account? If the mechanism is spore germination and exponential growth in the zone, risk should scale with cumulative time in the zone rather than with how often a dish was handled or reheated, and buffet outbreaks should be dominated by spore-formers rather than by the heat-labile pathogens of raw food. A body of buffet outbreaks driven mainly by organisms that cooking reliably destroys, with no relationship to holding time, would show this explanation is wrong.

• Why rapid cooling — shallow trays, blast chillers, two-stage cooling rules — is taken as seriously in commercial kitchens as cooking temperature.

Every term the collection defines is gathered in the glossary.